菜单
Primary review entrypoint for plans, diffs, commits, subagent results, and pre-submit checks. Selects quick, standard, or deep mode by scale and risk.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Planner-led repo improvement loops with subagents, review gates, Mermaid DAGs, and workflow-aware commits. Use for improve, refactor, audit, or continue-loop requests.
Triage and apply code review feedback with technical verification. Use when a human, reviewer subagent, CI, or adversarial reviewer returns findings or requested changes.
Heavy review mode normally invoked by request-code-review for risky checkpoints, pre-submit stacks, OpenSpec/cflx acceptance, security-sensitive work, or explicit deep review, audit, or red-team requests.
Post-implementation cleanup for recent changes using parallel reuse, quality, and efficiency reviewers. Use after building a feature or fix, before commit, PR, or submit.
End-of-workflow learning loop for skill-guided work. Use at session/checkpoint closeout to classify GitHub/CodeRabbit/human/CI findings, log missed-review patterns, and propose skill improvements only when evidence justifies them.
Use this skill when working in a repo that uses Graphite (the `gt` CLI) for stacked PRs. Activates whenever the git common dir contains `.graphite_repo_config`, the project's CLAUDE.md mentions Graphite or `graphite-pr`, or the user mentions `gt`, stacked PRs, atomic commits, or splitting a change into reviewable pieces. Drives the daily commit → submit → review → merge loop and the rules for sizing atomic commits. To bootstrap Graphite in a new repo, read `references/setup.md` first.
| name | request-code-review |
| description | Primary review entrypoint for plans, diffs, commits, subagent results, and pre-submit checks. Selects quick, standard, or deep mode by scale and risk. |
| version | 0.1.0 |
| author | George-RD |
| x-compatibility | jcode-primary, claude-code-adaptable, opencode-adaptable |
Use this skill as the main code review entrypoint. It can run a quick single-review, a multi-lens lightweight review, or escalate to deep-review for milestone/risky checkpoint review.
Core principle: review early, review precisely, scale reviewer effort to the change, and turn valid out-of-scope findings into follow-up tasks instead of expanding the current PR by default.
Default path: identify the target, load only relevant repo/domain skills or instructions, choose quick/standard/deep by scale and risk, dispatch the minimum useful reviewers, synthesize evidenced findings, fix in-scope blockers, and rerun gates.
HEAD~1..HEADbase..HEAD, often dev..HEAD, main..HEAD, or upstream branchdeep-reviewsubagent when bounded and cheap. Use swarm only when persistent coordination is needed.Read references only when needed:
references/review-rubric.md: severity definitions and checklist.references/reviewer-json-schema.md: machine-readable review output.references/adversarial-review.md: minimax or red-team prompts for adversarial/process/security lenses.references/multi-review-fanout.md: lightweight multi-lens roles and synthesis. Escalate to deep-review for milestone/risky checkpoints.assets/evidence-packet.md: copyable review packet template.scripts/validate.sh: behavioral contract checks for this skill.If a reviewer finds a real issue outside current scope, do not silently expand the PR. Create or propose a follow-up with evidence, impact, acceptance criteria, and priority unless it blocks correctness or acceptance for the current task.
Do not give the reviewer the whole session transcript unless the task is reviewing process or truthfulness. Focus review on plan/diff, requirements, and evidence. If reviewer feedback lacks concrete evidence or conflicts with project conventions, verify before changing code.